Hacker News Reader

Lots of dev opt for DBs like Redis or Mongo specifically because they're simpler and easier to develop for. Not for scaling. I think that can easily be a trap though because they are very easy for simple use cases, but get hairy really quickly when you start needing complex relationships.

View on HN

You can use RDBMS for anything that requires complex relationships and map them to unique identifiers in the Mongo documents.

In most cases though, Mongo fits use cases, we never try to use it as a wholesale RDBMS replacement.

Agreed... Mongo also works well for denormalizing data structures into something that renders faster to the browser. A single key/document->json->get->parse-> render is much faster than a query across many joins and/or sub-queries.

As an aside, if you're doing it anyway, it's also a good point to dump to .json.gz in S3 (or similar) as a secondary backup system.

Reply on news.ycombinator.com